Copyright 8 1995 by the Society of America

A P Element Containing Suppressor of Hairy-wing Binding Regions Has Novel Properties for Mutagenesis in Drosophila melanogmter

Robin R. Roseman,* Eric A. Johnson,+Chris K. Rodesch,: Michael Bjerke,” Rod N. Nagoshi: and Pamela K. Geyer*

*Department of Biochemist?, ?Genetics Program and :Department of Biology, The University of Iowa, Iowa City, Iowa 52242 Manuscript received March 22, 1995 Accepted for publication August 4, 1995

ABSTRACT P elements are widely used as insertional mutagens to tag genes, facilitating molecular cloning and analyses. We modified a P element so that it carried two copies of the suppessur of Hai?y-wing [su(Hw)] binding regions isolated from the gypsy . This transposon was mobilized, and the genetic consequences of its insertion were analyzed. Gene expression can be altered by the su(Hw) protein as a result of blocking the interaction between /silencer elements and their promoter. These effects can occur over long distances and are general. Therefore, a composite transposon (SUPor- P for 9presscPelement) combines the mutagenic efficacy of the gypsy element with the controllable transposition of Pelements. We show that, compared to standard Pelements,this composite transposon causes an expanded repertoire of and produces alleles that are suppressed by su(Hw) muta- tions. The large number of heterochromatic insertions obtained is unusual compared to other procedures, indicating that theSWor-Ptransposon may be useful for studying the structural and functional properties of heterochromatin.

LL organisms carry a wide variety of transposable ple P elements. Subsequent mapping, cloning and ge- A elements in their . InDrosophila, several netic analysis of a genetagged in this manner is compli- classes of transposable elements have been identified cated by the additionalPelements in the mutantstrain. which collectivelyconstitute 10%of the (RUBIN Mobilization of single P elements alleviates several of 1983). One of the best characterized of these is the P these problems. A large collection of mutant strains has element. Complete (autonomous) P elements are 2.9 been generated by this method (COOLEYet al. 1988a,b; kb in length and contain 31 bp inverted repeats that BERG and SPRADLINC1991; KARPEN and SPRADLING flank a unique region, which encodesa 1992). The utility ofthis technique to induce mutations activity (O’HAREand RUBIN 1983). in a given gene is limited by the low insertion rate of A great deal of research has elucidated both the cis P elements withinany single locus. However,as the and trans requirements forPelement mobilization. This benefits of isolation of mutations caused by single P research has allowed investigators to exploit Pelements insertions are great, methods are being developed to for several purposes, including germ line transforma- enhance the rateof single-element insertion into a par- tion (RUBINand SPRADLING1982) and isolation of ticular region of the chromosome (TOWERet al. 1993). genes by P element tagging (reviewed in ENGELS1988). P elements act as insertional mutagens. Mutations can Pelement insertions into or near a gene can be recog- result from interruption of the open reading fiame (Ru- nized in one of two ways. Integration of a P element BIN et al. 1982), interference with processing or stability may disrupt gene activity resulting in a mutant pheno- of the mRNA ( GEYERet al. 1988) or disruption of a regula- type, or the regulatory elements in a nearby gene may tory region (KELLEYet al. 1987) of a gene. However, inser- activate genetic markers carried by the transposon in a tions of P elements within or near a gene are notalways process known as “enhancer trapping” (O’KANEand associated with a phenotype. For example, three P ele- GEHIUNG1987; BEIRet al. 1989; BELLENet al. 1989; WIL ment insertions within the 5’ flanking region of the pour- SON et d. 1989). quoi$ms (wings-down) gene did not produce a mutantphe- Identification of genes by Pelement insertional muta- notype (SEGALATet al. 1992). These results suggest that genesis has been widely applied.In many instances, Pelements may not efficiently alter gene expression when mutations are induced by mobilization of several P ele- integrated in the proximity of upstream regulatory sites. As ments (KIDWELL 1986; ROBERTSONet al. 1988). In this P elements appear to insert preferentially into these way, mutations are generated in a backgroundof multi- control regions (KELLEY et al. 1987), it may be that alarge proportion of transpositions near genes go unrecog- Cmespondingauthm: Pamela K Geyer, Department of Biochemistry, nized. Therefore, increasing the ability of P elements to The University of Iowa, Iowa City, IA 52242. disrupt the action of enhancers should dramatically im- E-mail: [email protected] prove the insertional rate.

Genetics 141: 1061-1074 (November, 1995) 1062 R. R. Roseman et al.

A second well-characterized transposable element in gene. Mini-white contains 305 bp of 5' and 500 bp of 3' flank- Drosophila is the gypsy , which is 7.5 kb ing DNA and a deletion of most of the first (PIRROTTA 1988). The mini-whitegene was modified by insertion of 7uhit~ in length and contains long terminal repeats that flank regulatory sequences (-1084 and -1465 relative to the tran- a region encoding proteins homologous to retroviral scriptionstart site) upstream of -305 bp.This regulatory proteins (-,OR et al. 1986). These elements areeffec- region contains both the eye and testes enhancers (QIANrt tive mutagens since insertion of gypsy can disrupt gene (11. 1992) and directs a high level of white expression in the expression even when it resides as much as 10-80 kb eye.Two 430 bp su(Nw) binding regions (@sy sequences between nucleotides 647 and 1077, as numbered in MARLOR from a promoter (PEIFER andBENDER 1986; JACK et al. et al. 1986) flank the white gene to provide insulation of this 1991). Several tissue-specific mutations caused by the gene fromchromosomal position effects (ROSEMANet al. insertion ofgypsy have been identified (MODOLELLet 1993). Thesecond marker gene is the intronless yellow gene. al. 1983). In many cases, these mutations result from This yellow gene is 5.2 kb in length, contains 2.8 kb of 5' and the binding of the suppressor of Ha@-wing [su(Hw)] pro- 0.13 kb of 3' flanking DNA and lacks the tissue-specific bristle and tarsal claw enhancers (GMR and CORCES1987). tein to a region within the gypsy element near the 5' In experiment one, theSUPor-Pwas mobilized from the 60F (reviewed in CORCESand GEYER location as shown in Figure 1B. Transpositions were obtained 1991). The110-kD su(Hw) protein accumulates in most using a chromosomal source of transposase (P[ry+A2-?](99B) tissues throughoutdevelopment (PARKHURSTet al. ROBERTSONet al. 1988). Females carrying SUPnr-P were crossed to males that were w"'*; CyO/SP; Sb A2-?/TM6, Ubx. 1988; HARRISON et al. 1993). Association of the su(Hw) Single males carrying both SUPor-P and A2-? were mated in protein with its gypsy binding region causes inactivation individual vials to females that were y- nc- w"" or y'fs(l)Yb/ of enhancers when the su(Hw) binding region is placed FMO. New mobilizations were identified as straight-winged between an enhancer and its promoter (GEYERet al. (Cy') flies that were pigmented in the eye and body tissue 1986; HOLDRIDGEand DORSETT1991; JAGK et al. 1991). (Figure le).Stocks were established by backcrossing to either Effects of the entire gypsy element can be duplicated y- ac- w"" or y' fi(I)Yb/FMO. In the F2 generation, males were preferentially chosen to establish a stock. However, if by the su(Hw) binding region alone (HOLDRIDGE and the only y+ w+, Cy+ flies were females, then these individuals DORSETT1991; GEYERand CORCES1992). This binding were chosen. This undoubtedly biased our recovery of inser- region can generally inactivate both enhancers and si- tions on the X chromosome.As P element mobilization often lencers present in the Drosophila genome (ROSEMAN occurs premeiotically, each vial may have contained multiple et al. 1993). F2 progeny with an identical mobilization event. For this rea- son, only one line was established from each vial, unless flies The efficiency ofgypsy in disrupting 5' enhancer func- with different phenotypes were observed. tion may make it a more effective mutagen than a P The chromosomal location of each P element was deter- element. However, a major limitation for using gypsy as mined by a setof additional crosses in which males containing a mutagen is the inability to control its mobilization. In a new transposition were matedto y- nc- w"'x/y-uc-w"'x; this paper, we describe the use of a composite gypsy-P T(2,?)apx"/CyO, MKRS' females. (Cy0 andMKRS are balancer transposon as a mutator element. This transposon was chromosomes carrying the dominant Cy and Sb markers, re- spectively). X-linked insertions were identified because they created by cloningsequences corresponding to the did not produce y+w+ sons in this cross, while Y-linked inser- su(Hw) binding region between P element ends. We tions were identified because the only y+w+ progeny produced generated and characterized several hundred lines that were male. In addition, insertions that mapped to theYchro- carry a single composite element. We demonstrated mosome showed variegated pigmentation of the body and that this modified transposon combines the mutagenic eye, yellow7'"'whitr""'. Possible autosomal insertions were mapped by mating y+7u+; (30; MKRS male progeny to properties of gypsy and Pelements,resulting in a higher y-nc-7~"" females. The progeny from this cross were exam- mutagenic rate than found in standard Pmutagenesis ined to determine whether the y+w+ phenotype segregated and production of mutations that are dependent on away from either the second ( CyO) or third (MKR") chromo- the activity of the su(Hw) gene. In addition, we found some. that the composite gypsy-Pelement generates a substan- Virgin females and males that were phenotypically y+, w+, Cy and Sb were mated together to determine whether apartic- tial number of detectable insertions into or nearhetero- ular insertion had a recessive phenotype; between 150 to 200 chromatic regions. progeny were examined. In all but two cases, the insertion was classifiedas lethal if the cross generated none of the nonbalancer class of flies that was being scored. In the two MATERIALSAND METHODS exceptional cases, adult escapers were present, at a frequency Drosophila stocks: Flies were raised at 25"C, 70% humidity of less than 5% of the balancer class of flies. To determine on standard corn meal/agar medium. The mutations and whether a given insertionline was male or femalesterile, chromosomes used in this study are described in LINDSLEY between five and 10 homozygous SUPnr-P females and males and ZIMM (1992). were mated together. If that cross generated no progeny, an Mobilization of SUPor-P in M25 flies: The M25 flies carry a additional cross of either five to 10 homozygous SUPnr-1' fe- deletion of both the yellow and achaete genes (y- ac-), the males or males to five to 10 y-al--w"'x males or females was w///x mutation and thecomposite gypsy-Pelement inserted on carried out. Depending on theoutcome of these crosses, the the Cy0 balancer chromosome at cytological location 60F. A line was classified as male or female sterile. diagram of this transposon, called SUPor-P for %press%-E Lines in which the y+w+ markers did not segregate solely element, is shown in Figure 1A. This P element contains two from the X, Y, second or thirdchromosomes most likely con- marker genes. The first is a modification of the mini-while tained either fourth chromosome or multiple insertions. In Mutagenesis by a gypsy-P Element 1063 A.

PGP R BK PG ss K FIGURE1.-Mobilization of a composite gypsy-Pelementfor mutagenesis. (A) Structure of the mutator P element called SUPor-P. This transposon carries the yellow and white marker 1 kb genes. The direction of of each gene is indicated by the raised arrow. The P element ends areindicated by the small blocks at the end of the yellow and white genes. The 5' end of the P element is located at the 3' end of the white gene. Enhancers are shown B. by ovals; the yellow gene contains the wing (W) and body (B) enhancers; the white gene con- tains the eye (E) enhancer. Two su(Hw) bind- ing regions(triangles) flank the white gene. The restriction symbols shown below the line are as follows: B, BamHI; G, BglII; K, KpnI; P, PstI; R, EcoRI; S, SalI. (B)Genetic scheme used to generate new insertion sites of SWw-P. SUPor-P was mobilized from the Cy0 chromosome by P[ry+A2-3](99B), abbre- viated A2-3. Genetic symbols are described in 1 LINDSLEYand ZIMM (1992).

F2 SELECT y+ w+,non-CyO, Sb + males or females Will carry SUPor-P at new location

these cases, the number of insertion sites was determined by transposon in these lines is unknown, as they were lost prior Southern blot analysis that was designed to identify flanking to in situ localization. In summary, experiments two through restriction fragments. Multiple insertions gave a more com- five generated 77 lines that were analyzed for the associated plex restriction pattern than lines with one P element and recessive phenotype. Of these lines, 46 were also analyzed by were not considered further. Lines determined to contain a in situ hybridization. single insertion were categorized as fourth chromosomelines. Suppression of lethal second chromosome insertions by Mobilization of additional P elements containing su(Hw) mutations in su(Hw): The su(Hw) mutant stock used in these binding regions:Four additional small-scale Pelement mobili- studies was y-ac-7~~~ct'w'f'; Ro/ Cy0; b~'~'su(Hw)''/TM6, zations (experiments two through five) were carried out. In su(Hw)! Ubx. This combination of su(Hw) alleles reverses the experiment two,the starting SWor-P element was located in phenotypes associated with gypsy insertions and is female fer- the middle of 2L on a Cy0 balancer chromosome. This P tile. The su(Hw)" allele is a deletion of the su(Hw) gene (HAR- element was mobilized as described above (Figure le). RISON et al. 1992), whereas su(Hw)' is a point mutation in Twenty new mobilizations were obtained and mapped and one of the Zn fingers that retains some ability to bind DNA their associated recessive phenotypes were determined. Inex- (HARRISON et al. 1993). The Xchromosome in this stock car- periment three, the starting P element was SWw-P light (lt), ries two gypsyinduced alleles, ct6 f' that are suppressed by which is a derivative of SUPw-P that lacks the white eye en- su(Hw) mutations. This allowed identification of homozygous hancerbut is otherwise identical to SWw-P (see B.R. su(Hw) mutant flies. We studied only second chromosome >white>B.R. in ROSEMANet al. 1993). This P element was lines due to the fact that su(Hw) is located on the third chro- mobilized from position 92B on the TM6 chromosome by mosome. Crosses were designed such that no recombination crossing females carrying the SUPw-P It to w'"'; CyO/Sp; A2- could occur between the P-element-containing chromosome 3 Dr/TM6, Ubx males. Single males carrying both SUPor-P It and its homologue. Males carrying a lethal insertionof SUPw- and A2-3 were mated inindividual vials to y-~c-w"'~females. P were crossed to virgin females of the above su(Hw) stock. New mobilizations were identified as Dr flies that contained Males in the second generation that were phenotypically y+, pigmented eye and body tissues. Stocks were established by w+, ct, f and Rot were selected. These males were genotypi- backcrossing to y-~c-w"'~and selecting y+ w+; non-Dr flies. cally y-ac-w6' ct%'f'; SUPor-P/ CyO; su(Hw)" "I/+ and were We established 65 new lines, determined that 26 lines carried backcrossed to the su(Hw) stock. Males and virgin females insertions on the Cy0 balancer chromosome and analyzed that were phenotypically y+, w+, ct+, f+, Cy were selected. the remaining 39 for the associated recessive phenotype. Ex- These flies were homozygous for the su(Hw) mutations and periments four and five were mobilizations of SWwP ele- balanced for the lethal second chromosome SUPor-P inser- ments located on either the Xorsecond chromosome. These tion. Males and females of this genotype were mated, and at experiments generated 18 lines. The position of the starting least 100 Cy progeny were screened. Suppression of the SUPor- 1064 R. R. Roseman rt nl.

TABLE 1 Summary of recessive phenotypes generated by mobilization of SUPor-P in M25 flies

PercentagePercentage Percentage Percentage Percentage Percentage Chromosome Insertion.' Lethals" F. sterile" M. sterile" Visibld' "t I'

X 10 (62) I3 (8) 0 (0) 0 (0) 0 (0) 87 (54) 2 72 (432) I5 (65) 3 (13) 2 (9) 1 (3) 79 (342) 3 15 (88) 26 (23) 1 (1) 0 (0) 1 (1) 72 (63) 4 1.5 (9) ND ND ND ND ND Y 1 (6) 0 (0) NA 0 (0) 0 (0) 100 (6) Summary' .i88 16 (96)'i 2.5 (1 4)" I .5 (9)Il 1 (4)'l 79 (465)'' Totalinsertions (501

ND, not determined; NA, not applirahle; F, female; M, male; M't, wild type.Values in parentheses are number of lines. "Percentage of insertions on a given chromosome per total lines generated. "Percentage of insertions with a given phenotype per total number of insertions on a given chromosome. 'Summary of the number of lines in which recessive phenotypes were analyzed (excludes fourth chromo- some). "Percentage of insertions with a given phenotype per total number of chromosomes in the subset analyzed (no. phenotype/no. total insertions). This total exclutles the insertions on the fourth chronmsome.

P lethalphenotype resulted in thr production of straight- (1993) using roirit~DNA as a probe. This probe recognizes winged (Cy') progeny, which represented, in most cases, ap both the endogenous rohifr gene at 3C and the transposon. proximately one third of the homozygous .su(Hu~)flies. /"element positions were determined with resolution to the In situ hybridizations: Determination of the cytological loca- lettered intend. tion of SUPw-/'was clone according to the procedure of IAM Effects of the Y chromosome on variegation of SUPor-P: A

A. x4 X L x14L x2 x7NL x2 x1 6NL 2

3

B.

X x1L 902

2 25 30 35 40 -50 45 55 60

3 70 7565 70 95 8090 85 100 FIGURE2.-Chromosomal locations of SUPur-P. (A) Sites of insertion of SUPor-Pgenerated from the M25 line, which carried SUPor-Pat the 60F position. The location of SCJPor-/'elements associated with recessive lethal (lollipop, L) and nonlethal (bold line, NL) phenotypes are shown. In this study, f'element localization was carried out primarily on lines which had an associated recessive lethal phenotype. The number after the X indicates how many indepenclcnt lines had an insertion at that location. (B) Sites of insertion of .SUl"-/'generatcd by mobilizations from euchromatic positions. The locations of insertions associated with a recessive lethal (lollipop, I.) and nonlethal phenotype (bold line, NL) are shown. Eleven of the localized elements were generated from a starting position in the middle of 2% (experiment 2) and 35 of the localized elements were generated from a 92B starting position (experiment 3). These 46 lines were chosenat random, independent of' their associated recessive phenotype. Mutagenesis by a SrpSr-P Element 1065

TABLE 2 Complementation analysis of SUPor-P-induced mutations

Line Insertion site Allele or todeficiencyFailure tested complement 546-1" 7B 546-1" Ct6 Yes 203-1" 33F Df(2L)al: 21B8-Cl; 21C8-D1, 2201-2; No 33F5-34Al 2742" 55BC Df(2R)PC4: 55A-F No 41 1-1" 55D Df(2R)PC4: 55A-F Yes 4251" 55D Df(2R)PC4: 55A-F Yes 286-4" 61D Df(3L) emz, red': 61C3-C4; 6248 Yes 82c" 73CF Df3L)std1', Ki roe p[p]: 730173Bl; Yes 1614 77A stDf(3L)rdg th C, in ri p[p]: 77A1; 7701 Yes 850-1" 85A Df(3R)CA3: 84F2; 85A5-7 Yes 59-2" 87E Df(3R) ry nZ7,cu Dfd dar: 8701-2;87Fl-2 Yes 126M19' lOOCD Df(3R)awd,lOOGD KRB, ca: Yes "SUPm-Plines generated from the 60F position. 'SUPm-P lines generated from a non-6OF position. large percentage of mobilizations showed variegationof the RESULTS marker genes, yellow and white. To determine whether this Mobilization of a Pelement containing su(Hw) bind- variegation was sensitive to modifiers of position effect varie- gation, we tested the effects of changing the Y chromosome ing regions: The parental M25 line is a y- w- strain that dosage. Females carrying SUPm-P weresrossed to males that carries a SWw-P elementinserted on the Cy0 chromo- carried an attached XYchromosome [ XY, y VI. All male prog- some at cytological location 60F (Figure 1).The SWw- eny from this cross were XO. The phenotype of these males P transposon carries modified yellow and white genes was compared with that of males obtained from crossing SU- which confer a red/brown eye color and dark pigmen- Pm-Pcontaining females to y-ac-w"'*. Position effectvariega- tion is very sensitive toculture conditions such as temperature tation of only body and wing cuticle. Transpositions of and media. For this reason, these crosses were carried out SWw-Pfrom the Cy0 balancer to another chromosome under identical conditions at 25". Similar crosses carried out were identified as straight-winged flies that showed w+ with nonvariegating SUPm-Plines showedthat these lineswere and/or y+ expression. Altogether, 1759 fertile crosses not affected by Y chromosome dosage. were established and 655 independent lines were ob Effects of the su(Hw) protein on marker gene expression: Several centric and telomeric insertion lines were obtained. tained. Our "jumpingrate," defined as the percentage To determine whether the presence of the su(Hw) protein of vialscontaining at least one y+ w', Cy' fly (BERGand provided protection from repressive position effects that are SPRADLINC1991), was 37%. In most cases, the pheno- associated with these regionsof the genome, lines containing type of fliescarrying the mobilized SWw-P elementwas these SUPm-P transposons were made homozygousmutant for indistinguishable from the parental line. However, a su(Hw). These crosses were carried out as described for the suppression studies. The resulting phenotypes were classified number offlies showed phenotypes that werevarie- by considering both the level of marker gene expression and gated either for both thewhite and yellow genes, or only the amount of variegation observed. for the yellow gene.

TABLE 3 Chromosomal distribution of variegating lines obtained by mobilizationof SuPor P in M25 flies

Chromosome Percentage Insertions Class Class 11' Class III~

x" 8 (29) 100 (29) 0 (0) 0 (0) 2" 73 (255) 82 (209) 7 (18) 11 (28) P (52) 15 77 (40) 4 (2) 19 (10) 4" 3 (9) 0 (0) 100 (9) 0 (0) k" 1 (4) 0 (0) 100 (4) 0 (0) Summaryb 80 (278) 9.5 (33) 10.5 (38) Total" 349 Values in parentheses are number values. " Percentage of insertions with a given phenotype per total number of insertions on a given chromosome. Percentage of insertions with a given phenotype per total inserts analyzed. A subset of the total transposition lines were analyzed in this study. Wild-type eye and cuticle phenotype. "Variegated phenotype of both eye and cuticle. /Variegated phenotype of cuticle only. 1066 R. R. Roseman et al.

A.

302-1 2 chromocenter

690-1 X

450-1 2R te lornere

4 4a 3L te lornere

31 6-1 3 R te lornere

755-1 4 telomere

FIGURE3.-Phenotypes of marker genes associated with 3uhPinsertions. (A) Eye phenotypes or male flies carrying a SUPor- P transposon integrated ateither a centric or a telomeric location. Three genotypic backgrounds are shown for each insertion. (B) Body phenotypes of males carrying a SUPor-P transposon integrated at the same positions as in A.

Of the 655 lines, 601 were analyzed further (Table insertions. Second,Ptransposition events occur primar- 1). In our mutagenesis protocol, the mobilization of ily in premeiotic germ cells. Thus,if the X-linked inser- SWw-P was induced in males. New insertions could tion caused inactivation of a cell autonomous gene es- be recovered on either one X, the Y, one second (a sential for viability or fertility, this germ cell would fail transposition in cis on the Cy0 chromosome would be to form afunctional sperm. In contrast to these predic- lost in our scheme),one third (the A2-3 chromosome tions, we found that most lines had insertions on the was selected against), or two fourth chromosomes. homologoussecond chromosome (72%), withonly Based on this analysis,we anticipated that there should 15% of the lines containing an insertion on the third be an equal number of insertions recovered on the chromosome. A low percentage of insertions were re- second and third chromosomes. The frequency of X- covered on the other chromosomes: X (lo%), Y (1%) linked insertions was expected to be reduced for two and fourth (1.5%) (Table 1). reasons. First, F2 males were usually used to establish Among these SWw-P lines, a second bias was ob- stocks which biased against the isolation ofX-linked served. Only15% of the second chromosome insertions Mutagenesis by a Srpsr-P Element 1067

B. XY, su(Hw)+ X3,su(Hw) ' XY, su(Hw)-

302-1 2 chrornocente

690-1 X telomere

450-1 2R telomere

4 4a 3L telomere

31 6-1 3R te lomere

755-1 4 telomere

FIGURE3.- Continued were associated with a recessive lethal phenotype, in due to inherent properties of SWw-P, but is caused by contrast to 26% for those on the third chromosome. its starting location in the parental M25 line (this will In other mutagenic screens involving singlePelements, be examined in detail in a later section). For these the rate of recessive lethality of new insertions ranged reasons, we will not consider the 60F mutations in the from 10 to 17% for the second chromosome and 10 to following characterization of SWor-P-induced lesions. 12% for the third chromosome (COOLEYet al. 1988a,b; Complementation analysis of gytsy-Pelement muta- BEIR et al. 1989; COOLEYet al. 1988a; KARPEN and tions: Insertion of a P transposon causes the majority SPRADLING1992). Therefore, our lethality rate on the of mutations isolated in single Pelement mutageneses second was similar to previous studies, while that on (COOLEYet al. 1988b; BEIR et al. 1989; BERG and the third chromosome was about 2.5-fold higher. These SPRADLING 1989;ZHANG and SPRADLING,1994). Based differencesbetween the second and third chromo- on the following data, we believe the same istrue for the somes reflectthe disproportionate recovery of linesthat SWor-P. The location of SWm" in several randomly carry SWm" in a 60F location, the same cytological chosen recessive lethal lines was determined by in situ position as the parental SWw-P insertion (see below). hybridization (Figure 2). We then investigated whether Subsequent analyses indicated that this anomaly is not any of our SWw-P-induced mutations were uncovered 1068 R. R. Roseman et al. by known cytologicaldeletions of these regions. In eight tions were located on the Y (4/33) and fourth (9/33) of 10 lines tested, the location of the mutations corres- chromosomes, two chromosomes which arehetero- ponded to the SUPor-Pinsertion site, as determined by chromatic in nature.The Pelementwas localized in six failure to complement available deletions in the region of the nine lines designated to the fourth chromosome of the transposon (Table 2). In addition, one allele of by genetic mappingstudies. In all cases,the fourth chro- cut (ct) was recovered that carried a SUPw-P insertion mosome assignment was confirmed (Figure 2A). Al- at the appropriate cytological location, 7B. Thus, we though insertions into these heterochromatic chromo- find a strong correlation between the location of a le- somes caused both yellow and whitevariegation,the level sion and the site of SlJPor-Pinsertion for several muta- of expression from these reporter genes was dramati- genic events. cally different. Three of the four Y chromosome inser- A subset of SUPw-P mutations are suppressed in a tions had greatly reduced expression of both marker su(Hw) mutant background A potential advantage of genes; eye color was essentially white with a few spots the SWor-P transposon as a mutagen is the generation of pigment, and cuticle color was essentially yellow with of lesions that are dependent on the presence of the a few spots of color. In contrast, lines carrying an inser- su(Hw) protein. We predicted that asubset of the SUPor- tion in the telomereof the fourth chromosomeshowed P-induced mutations should be suppressed in a su(Hw) nearly normal levels of white and yellow gene expression mutant background. Forty-one randomly selected sec- and only light variegation (Figure 3). In theonly fourth ond chromosome lines carrying recessive lethal P ele- chromosome line localized to centric heterochromatin, ment insertions were crossed into a SU(HW)"/SU(HW)' pigmentation of the eye,body and wing cuticle was mutant background. In nine cases (22%), the lethality greatly reduced and showed extreme variegation. was reversed. The fact that the lethal phenotypeof these The remaining yelloz~""~,white""' insertions (20/33) mutations was sensitive to the allelic state ofsu(Hw) mapped to the second or third chromosome. In these correlates the mutant phenotype with the presence of cases, the level of marker geneexpression varied greatly the su(Hw) binding regions. This indicates that these among lines, with most lines showing low to intermedi- insertions would not produce a detectable phenotype ate levelsof pigmentation and moderate variegation if a standard P element was inserted at the same site. (Figure 3). The SUPor-P element associated with ten of A high percentageof transposons show chromosomal the yello~'"~,white""' autosomal lines (50%, 10/20) was position effects The SUPor-P transposon carries two localized by in situ hybridization and foundto reside in visible, cellautonomous phenotypic markers, yellow and the centric region of the second or third chromosome white. We noted thatmany lines generated from mobili- (Figure 2A). zation of the M25 SUPw-Pdid not express these markers The Y chromosome is a strong modifier of position in a wild type pattern. A random subset of 349 lines was effect variegation, with loss of the Y chromosome re- characterized in moredetail. Flies carrying this transpo- sulting in increased variegation (SPOFFORD1976). We son could be classified into three phenotypic classes tested whether the autosomal and fourth chromosome (Table3, Figure 3). The predominant class I (80%, yellow""', white""' lines were sensitive to the dosage of the 278/349) had a marker phenotype identical to the pa- Y chromosome. In these studies, females carrying the rental M25 strain, showing dark eye, body and wing SWor-P transposon were mated to attached XY males. color. Class I1 (9.5%, 33/349) showed avariegated phe- All male progeny generated in this cross were XO. The notype in boththe cuticle and eye: yellow""',white""'. level of pigmentation associated with the eye and body Class I11 (10.5%, 38/349)showed only a variegated cuti- cuticle was compared with males generated in a parallel cle coloration: yellow""'. Variegated phenotypes are of- cross between SUPor-P females and y-ac-w'''*males. ten observed when euchromatic genes are placed next This allowed comparison of males with different doses to or within heterochromatic portions of the chromo- of Y chromosomes. We tested sixteen yellow""', white""' somes, such as the centromere or telomere (SPOFFORD second chromosome lines, two yellow""r,white""' third 1976; HENIKOFF 1990). Theseposition effects are pro- chromosome lines and six yello~""~,white"'" fourth chro- posed to result from a variable spreading of highlycom- mosome lines for effects of the Y chromosome. In all pacted , from theheterochromatic region cases, XY males showed substantially less variegation in into euchromatic DNA (LOCKEet al. 1988) or compart- the eye than X0males. One caveat of this experiment mentalization of heterochromatic genes into regions of is that the genetic background of the X0and XY males the nucleus that areinaccessible to transcription factors differs, which may affect the observed phenotype. Posi- (WAKIMOTOand HEARN1990; DEVLINet al. 1990). Gene tion effect variegation is sensitive to a large number of expression is repressed if the euchromatic gene is in a modifiers that could differ in these backgrounds. If the heterochromatic region,whereas it remains active when enhanced eye phenotype in the X0 background was in a euchromatic compartment. due to the presence of dominant modifier mutations Our chromosome mapping and in situ localization and was not related to the loss of the Y chromosome, data provided insights into the origin of these pheno- then we would have expected to see substantial pheno- types (Table 3, Figure 2). Several of the class I1 inser- typic variations in the eyes among the male progeny of Mutagenesis by a gypsy-P Element 1069 a given line. As this was not the case, these results sug- 4, Figure 3). These experiments aresubject to the same gest that the position effects on white expression can caveatsas applied in the previous experiments. We be modified by the dosage of the Y chromosome (Fig- found that loss of the Y chromosome failed to enhance ure 3A). the white phenotype. Furthermore, theyellow phenotype The effects of the Y chromosome dosage on yellow was enhanced in only four of the 11 lines (36%).Again, expression dependedupon the chromosomeinto substantial phenotypic variations in the cuticle color which the SUPor-P element was inserted. Second and were not observed among the male progeny of a given thirdchromosome yellow""',white"" lines showed less line, suggesting that in lines where yellow expression yellow variegation in XY males than XO, while yellow was affected, this was caused by the dosage of the Y expression in the fourth chromosome SWm-P inser- chromosome. Thesestudies indicate that the Ychromo- tions did not change in these two backgrounds (Figure some does not act as a general enhancer of telomeric 3B). As with white, the X0males produced from a given position effects for insertions located on the X, second line did not show substantial phenotypic variation, sup- or third chromosomes. porting the hypothesis that these effects are associated Differentialrecovery of mutationson the auto- with changes in dosage of the Y chromosome. somes: As noted previously, there was a bias in the re- Altogether, 23 of the 33yellow""', white""' SUPor-Plines coveryof SWor-P insertions on the second chromo- were found to be in inserted in heterochromatin. In 13 some. A series of experiments were done to examine lines, SUPor-P resided inheterochromatic chromo- the substantial contrasts in the rates of insertion and somes (fourth and I? ; in 10 lines, SWor-Pwas localized recessive lethality between the second and third chro- by in situ hybridization to centric regions of the second mosomes. The distribution of the SUPor-P elements in or third chromosome. These results suggest that the 21of the 23 third chromosome recessive lethal lines yellow""', white"' phenotype is an indicator of whether was determined to test whether an insertional hot spot an insertion of SWm-Pis in or near heterochromatin. on this chromosome caused the increased mutational The position of the SUPor-Pelement in approximately rate. In 18 of these 21 cases, the transposon associated one third (13/38) of the class I11 lines was determined with a recessive lethal phenotype was localized to a dif- (Figure 2A). Twelve lines had SWmP transposons in- ferent chromosomal position (Figure 2A). Only one serted at a telomere of one of the large chromosomes. cluster of four insertions was detected, which occurred These regions of the genome have been shown pre- in the telomere of X,cytological position 100F. This viously to repress expression of genes (HAZELRIGG et al. indicates that there is not a strong insertional hot spot 1984; LEVIS et al. 1985; KARPEN and SPRADLING1992; of recessive lethal mutations on the third chromosome. WALLRATHand ELGIN, 1995). Although, most of the However, there appears to be a general clustering of class I11 insertions were located at a telomere, other SWor-P transposons at the ?R telomere, as noted by lines shown to contain telomeric insertions either did the localization of SUPor-Pin nine additional nonlethal not show any position effects on yellow expression or lines (Figure 2A). showed reduced but not variegated yellow expression We also tested whether a preexisting lethal was pres- and were not included in class 111. This suggests that ent that would increase the mutation rate on the third either thesu(Hw) binding regions associated with SUPor- chromosome. Background lesions have affected origi- P protected both the white and yellow reporter genes nal estimations of mutationrate in similar studies from chromosomal position effects or that SUPor-P was (COOLEYet al. 1988b; BELLENet al. 1989). Complemen- not inserted into sequences that caused repression of tation analysis was done with 10 of the lethal third chro- gene expression. To distinguish between these possibili- mosome lines (10/23). Each of these lines was crossed ties, we crossed eleven telomeric insertion lines into a to the remaining lines in allpair-wise combinations. su(Hw) mutantbackground (Table 4, Figure 3). We Complementationoccurred in every case, indicating found that the eye phenotype associated with all lines thatthe lethals representedindependent, separable became much lighter and variegated in a su(Hw) mu- mutations. We concludethat for the third chromo- tant background. This demonstrates that the su(Hw) some, the recessive lethals generated were distributed protein is effective at protecting white from repressive at a number of different sites. chromatin assembled at chromosome tips. The yellow The same conclusion could not be drawn when the phenotype associated with several lines also changed second chromosome recessive lethal lines were ana- when placed in a su(Hw) mutant background, although lyzed in detail. Localization of the SWm-P transposon this effect was less consistent (Table 4, Figure 3B). in 27 lines associated with a second chromosome reces- We examinedwhether the phenotypes associated sive lethal mutation showed that 14 contained an in- with telomeric insertions were sensitive to the dose of sertionat the telomere of 2R, cytological position the Ychromosome, as was demonstrated for the hetero- 60F, as did 80% (16/20) of a randomly selected group chromatic insertions in the class I1 transposons. Females of nonlethal, nonvariegating second chromosomelines from several different telomeric insertion lines were (Figure 2A). Complementation analysis was done be- mated to attached XY males as described above (Table tween 11 of the 14 60F recessive lethal lines. Each of 1070 R. R. Roseman et al.

TABLE 4 Phenotypes of marker genes in lines carrying telomeric insertions of SuPur-P

yellow phenotype white phenotype No. location su(Hu~)+,XY su(Hw)+, X0 su(Hru)-, XY su(Hw)+,XY ~u(Hw)+,X0 su(Hw)-, XY 690-1, 1A black black brn, slt var red red or, hvy var 863-1, 1A black black brn, slt var red red or, md var 178-1, 60F brn brn brn, slt var red red or, ext var 315-1, 60F dk brn brn dk brn red red or, hvy var 450-1, 60F brn, md var brn, md var brn, md var red red or, ext var 50-2B, 60F It brn, md var It brn, md var It brn, md var brn brn or, ext var 44a, 61A It brn, md var It brn, hvy var It brn, md var red red w, ext var 108-1, 1OOF black black tan, md var red red w, ext var 316-1, l00F It brn, md var tan, ext var tan, md var red red w, ext var 478-1, l00F brn, md var tan, hvy var brn, md var red red w, ext var 525-1, lOOF brn, md var It brn, md var brn, hvy var red red w, ext var

~~~ ~ ~ ~ ~~ It, light; dk, dark; var, variegation;slt, slight; md, medium; ext, extreme; hvy, heavy; brn, brown; or, orange; w, white. The level of yelloru gene expression produced flies with pigmentation in the descending order: black > dark brown > brown > light brown > tan > yellow. The level of white gene expression produced flies with eye color in the descending order: red > brown > raspberry > orange > yellow > white. Variegation was classified in descending order: no variegation > slight variegation > medium variegation > heavy variegation > extreme variegation. these 11 lines was crossed to the remaining lines in all recessive lethal mutations: We determined the reces- pair-wise combinations. Complementation occurred in sive phenotypes associated with 77 linesgenerated every case, indicating thata preexisting second chromo- from these small-scale mobilizations (experiments two some lethal was not present and that the lethal mu- through five, Table 5). In this case, the recessive lethal tations at this position resulted from multipleand sepa- phenotype associated with second chromosome inser- rable lesions. This latter conclusionwas further demon- tions was 30% (14/46), while the third chromosome strated by complementation tests using a deletion in recessive lethality frequency was found to be 25% the 60F region (kindly providedby N. PATEI.).Thirteen (5/20). These frequencies are very similar to that of lethal 60F alleles were tested,and fourfailed to comple- the third chromosome lethal rate found in the original ment this deficiency.We conclude that the secondchrc- M25 mutagenesis (26% or 23/88). mosome lines represent the biased recovery of lines Taken together, these data suggest that the reduced with an insertion at cytological location 60F. frequency of mutationscaused by presumed second SUPor-Pelements mobilized from other locations are chromosome insertions in theM25 mutagenesis derives not targeted to 60F Inclusion of somesequences from the 60F insertion class. If we discount this second within a P-element vector can alter the specificity of P- chromosome data due to this anomaly,we find that the element insertion (KASSIS et al. 1992). To verify that the frequency of recessive lethal mutations derived from large number of 60F insertions obtained frommobiliza- the M25 third chromosomeevents and both the second tion of the M25 SUPor-P element was not influencedby and third chromosomes from the four subsequent mu- a general targeting property of P elements carrying a tageneses is approximately 27% (42/154). su(Hzu) binding region, small-scale mobilizations with four differentSUPor-Por SUPor-P It (a modified element DISCUSSION lacking the white eye enhancer) transposons were car- ried out (experiments two through five). Forty-six ran- SlJPor-Pmutagenesis: These studies examined whether domly selected lines were mapped by in situ hybridiza- addition of su(Hw) binding regions to a marked P ele- tion, and none contained an insertion at 60F (Figure ment increased the effectiveness of singleP-element 3B). We conclude that the clustering of SUPor-P ele- mutagenesis.This was predictedbecause the su(Hw) ments in cytological position 60F that were recovered binding region is a potent mutagen of 5’ regulatory in the M25 study was a consequence of the starting regions (GEYERet al. 1986; HOLDRIDGEand DOR~ETT location of the P element. In the presence of a source 1991;JACK et al. 1991; ROSEMANet al. 1993), unlike other of transposase, imprecise male recombination between P transposons (SEGALATet al. 1992). Several mobiliza- the telomere of the parental M25 chromosome and its tion experiments were carried out using the SUPor-P homologue would give results consistent with this bi- transposon. As a measure of the mutageniceffectiveness ased recovery. of this transposon, we determined the frequencyof pro- SUPor-P elements cause a high rate of production of duction of recessive lethal mutations. We found that Mutagenesis by a ~psy-PElement 1071

TABLE 5 Summary of recessive phenotypes generated by mobilization of composite gypsy-P elements not located at 60F

Percentage Percentage Percentage Percentage Chromosome insertions" lethal" F. sterile6 Wt'

X 8 (6) 0 (0) 0 (0) 100 (6) 2 60 (46) 30 (14) 9 (4) 57 (26) ? 26 (20) 25 (5) 0 (0) 75 (15) 4 4 (3) ND ND ND Y 2 (2) 0 (0) NA 100 (3) Summary' 74 26 (19)" 5 (4)" 66 (49)" Total 77

~ F, female; M, male; Wt, wild type; N, number; ND, not determined;NA, not applicable. Values in parentheses are number values. Percentage of insertions on a given chromosome per total lines generated. "Percentage of insertions with a given phenotype per total number of insertions on a given chromosome. 'Refers to the number of lines in which recessive phenotypes were analyzed (excludes the fourth chromo- some). 'The percentage of insertions with a given phenotype per total number of chromosomes in the subset analyzed.

27% of the SUPor-Plines wereassociated with a recessive vation of control elements required for some genes. lethal phenotype (42/154; 95%confidence interval of Given this caveat, it is difficult to predict what percent- 20 to 34%). This compares with a frequency of 12% age of alleles should be suppressible. However, the cre- (958/7825; 95%confidence interval of 11.5 to 13%) ation of P-element-induced alleles that are suppressed for recessive lethal mutations on the second and third by mutations in the su(H7u) locus demonstratesthat chromosomes in the large mobilization screen reported SlPor-P can expand the repertoire of mutations made by KARPEN and SPRADLING(1992), which is representa- by P mutagenesis. tive of the frequency obtainedin other studies (COOLEY The SUPw-P transposonmay provide a means to re- et al. 1988a,b;BIER et al. 1989).These data indicate that cover insertions into regions of the genome not ob- mobilization of SUPor-P caused a significant increase in tained by standard P mutagenesis: P elements appear the frequency of recessive lethal mutations. to insert preferentially intoeuchromatic positions A novel class of P-induced alleles are generated by (ENGELS1988; ZHANG and SPRADLINC1993; WALLRATH insertion of SUPw-P: The increased frequency of muta- and ELGIN,1995). The low recovery of marked P ele- genesis is coincident with the production of alleles of ments into heterochromatic regions has been proposed SUPor-P that were suppressed by su(Hw) mutations. We to result from compromised detection (ENGELS 1988; found that 22% (9/41)of the second chromosome re- KARPEN and SPRADLINC1992; ZHANC and SPRADLING cessive lethal mutations were viable in a su(Hw) mutant 1994).In fact, P-element mobilization under conditions background, suggesting that bindingof the su(Hw) pro- that suppress position effect variegation allows recovery tein to SWor-P caused the mutant phenotype. These of insertions intoheterochromatin (ZHANG and suppressible alleles are unlikely to result from insertion SPRADLJNG1994). Mobilization of the M25 SUPor-P ele- of gypsy elements because new excisions or insertions ment generated many lines that carried insertions into of gypsy do notoccur under hybrid dysgenic conditions heterochromatic regions of thegenome. Although (WOODRUFFet al. 1987; EGGLESTONet al. 1988). The these lines show variegation of both yellow and white, 22% frequency of suppressible alleles is lower than was the presence of the su(Hw) binding regions greatly re- expected considering the 2.5-fold increase in mutage- duced the negative effects of these regions. In a su(Hw) nicity. One possible explanation for theobserved lower mutant background, variegation became more severe; rate is that these suppression studies used a hypomor- and, in some cases, pigmentation was barely detectable phic combination of su(Hw) alleles because su(Hw) null (Table 4, Figure 3). The ability to isolate transposons mutantsare female sterile. This allelic combination inserted into heterochromatin appears to be a general completely suppresses somatic phenotypes caused by feature of composite gypsy-P elements, as our smaller gypsy insertion but produces sufficient amounts of pro- mobilizations (experiments two through five) produced tein in the germline to allow fertility. Thus, it is possible 16% (12/77)of the lines showing yellow and whitevarie- that su(Hw)protein made in the germlineis transported gation (Table 5). These results support the notion that into the oocyte and remains at sufficient levels in the position effects limit the recovery ofmarked Pelements developing embryo to bind to SWor-Pand cause inacti- into heterochromatin. Furthermore, they suggest that 1072 R. R. Roseman et al. a second method for obtaining large a number of inser- models have been proposed to explain how heterochro- tions into this region of the genome is the mobilization matin represses euchromaticgene expression. Most of P elements carrying su(Hw) binding regions flanking models evoke inactivation of transcription as the under- a marker gene. lyingbasis for repression. The chromatin assembly Chromatin assembled at and centromeres model proposes that transcriptional inactivation results show different effects on expression of marker genes from chromatin compaction due to assembly of large in SUPor-P: From the mobilization of the 60F SWor-P, protein complexes that associate into a higher order many transposons located at telomeric positions were structures (SPOFFORD1976; LOCKEet al. 1988). In this obtained. In total, 18 lines had insertions on the X, 3L model,chromatin compaction initiates at particular and 3R tips (Figure 2A). An additional 30 transposons sites and spreads along the chromosome; the closer a were associated with the 2R telomere. Although these gene to the site of initiation, the morelikely repression 60F insertions may have resulted from male recombina- of gene expression will occur. Studies on chromosomal tion, they differ in the expression of the marker genes rearrangements that juxtapose euchromatin and het- and therefore were useful to study telomeric position erochromatin provide some support forthis hypothesis. effects (Table 4). In many rearrangements, inactivation of euchromatic The level of white expression was wild type or nearly genes occurs over long distances with the gradient of wild type at all telomeric sites; whereas, there was range gene inactivation matching the change in morphology of yellow phenotypes, with some lines showing wild type of the euchromatic region towards a more heterochro- pigmentation and others showing a low level of varie- matic appearance (SPOFFORD1976). In this context, the gated coloration (Table 4, Figure 3). The phenotypes su(Hw) protein may block repression by heterochroma- associated with these marker genes reflect the fact that tin by providing a road block to the subsequent spread- white is completely flanked by su(Hw) binding regions, ing of repressive chromatin, eitherby its direct interac- while yellow is not. This suggests that unlike heterochro- tion with DNA or as a consequence of establishing a matic position effects of the centromeres, fourth and Y domain boundary. chromosomes, telomeric position effects can be com- A second model is the compartmentalization model pletely blocked by the su(Hw) protein. This observation (WAKIMOTOand HEARN1990; -EN 1994). In this that su(Hw) protein differentially insulates the white model, heterochromatin and euchromatin occupy dif- gene from repression in these genomic locations sug- ferent regions of the nucleus. It is envisioned that in gests that telomeric position effects are distinct. Addi- theseregions transcription factors required for ex- tionally, the expression of marker genes present on pression of genes are distinct and that the two nuclear SUPor-P elements inserted into centric regions and the compartments are differentially enriched with the ap- fourth chromosome was sensitive to the Ychromosome propriate factors. Thus, genes requiring euchromatic dose, while expression of these genes was affected only transcription factors would be poorly expressed if local- in a minority of telomeric insertions on thesecond and ized into the heterochromatic nuclear compartment. third chromosomes. This second observation supports This model is supported by the studies on the effects the idea that chromatin assembled at these telomeres of chromosomal rearrangements on expression of het- has different propertiesfrom centric or fourth chromo- erochromatic genes (WAKIMOTOand HEARN1990). In some chromatin. These data are consistent with those this context, it is possible that thesu(Hw) protein blocks Of WALLRATHand ELCIN (1995),who found thatexpres- heterochromatic position effects by establishing an in- sion of the hsp 70-white gene present on P elements dependent domain that encompasses the euchromatic inserted into telomeres of the second and third chro- gene and alters the compartment into which this gene mosomes was not altered by known modifiers of posi- is localized, thereby increasing its expression. tion effect variegation, including the Y chromosome. Both models described above are based on data oh The difference between repressive centric and telo- tained from studies of centric position effects. It is un- meric chromatin may indicate that a uniqueset of pro- clear whether these models adequately describe telo- teins are involved in establishing these kinds of chro- meric effects. Analysisof the SUPor-Pinsertion lines may matin. Alternatively,this may reflect the fact that provide insights into this question. In this regard, we establishment of centricheterochromatin requires a found that several telomeric insertion lines showed dis- greater amount of each protein than is needed for te- parate effects on the expression of the SUPor-P reporter lomeric chromatin. Thus, if proteins which establish genes. In these cases, white expression waswild type, repressive chromatin interfere with the association of while yellow expression was low and/or extremely varie- the su(Hw) protein, then the binding of this protein gated. This wild-type level of white expression required may bemore difficult incentric positions than te- insulation by the su(Hw) protein (Table 4, Figure 3). lomeric, producing a less effective block at these posi- Thus, if the telomeric position effects resulted from tions. inappropriate compartmentalization in the nucleus and The mechanism by which the su(Hw) protein insu- insulation by the su(Hw) protein occurred by altering lates the mini-whitegene from positioneffects: Several the nuclear compartment of the telomeric transposon, Mutagenesis by a SrpSr-P Element 1073 we would have expected that theyellow gene would also ture media. We thank NIPAM PATELfor sending the 60F deficiency localize to the new compartment, and its gene expres- chromosome. This work was supported by NationalInstitutes of Health grant GM-42539 to P.K.G. and grant GM-45843 to R.N.N. sion would be largely unaffected. However, this is not observed. Alternatively, if telomeric repression resulted LITERATURE CITED from the assembly of an altered, perhaps more com- pactedchromatin structure, and insulation by the BELLEN,H. J., C. J. O'KANE,C. WILSON,W. GROSSNIKLAUS,R. K. su(Hw) protein occurred by blocking the assembly of PEARSONet al., 1989 P-element-mediated enhancerdetection: a versatile method to study development in Drosophila. Genes the white gene into this repressive structure, then we Dev. 3: 1288-1300. would predict that yellow expression would be suscepti- BERG,C. A,, and A. C. SPRADLING,1991 Studies on the rate andsite- ble to repression, since this gene is not flanked by specificity of P element transposition. Genetics 127: 515-524. BIER,E., H. VAESSIN,S. SHEPHERD,K. LEE, K. McCALL et al., 1989 su(Hw) binding regions. This model is more consistent Searching for pattern and mutation in the Drosophila genome with our results and suggests that telomeric repression with a P-lac 2 vector. Genes Dev. 3 1273-1287. may not becaused from inappropriate compartmental- COOI.EY,L., C. BERG,and A.C. SPRADLING,1988a Controlling P element insertional mutagenesis. Trends Genet. 4 254-258. ization. However, it is possible that the disparate yellow COOLEY,L., R. KELLEY and A. C. SPRADLING,1988b Insertional muta- and white phenotypes are due to the fact that these genesis of the Drosophila genome with single P elements. Sci- genes are expressed in distinct cell types or that the ence 239: 1121-1128. CORCES,V. G., and P. R GEYER,1991 Interactions of retrotranspo- threshold amount of each product required to elicit a sons with the host genome: the case of the gypsy element of phenotype differs. Thus,more experiments are re- Drosophila. Trends Genet. 7: 86-90. quired to sort out these possibilities. DEVI.IN,R. H., B. BINCHAM andB. T. WAKIMOTO,1990 The organiza- tion and expression of the light gene, a heterochromatic gene In yeast, telomeric position effects are caused by the of Drosophila melanogaster. Genetics 125 129-140. assemblyof an altered chromatin structure which is EGGLESTON,W. B., D. M. JOHNSON-SCHLITZand W. R. ENGEIS,1988 initiated at the endof the chromosomeand propagated P-M hybrid dysgenesis does not mobilize other transposable ele- ment families in D. melanogaster. Nature 331: 368-370. a short distance in the proximal direction (RENAULD ENGELS,W. R., 1988 Pelements in Drosophila, pp. 437-484 in Mo- et al. 1993). Interestingly, in many SWrn-P telomeric bile DNA, edited byD. BERGand M. HOW, ASM Publications, insertion lines, the yellow gene was partially protected Washington, DC. GEYER, P.K., and V. G. CORCES,1987 Separate regulatoty elements from repression by the su(Hw) protein (Table 4). These are responsible for the complex pattern of tissue-specific and results may reflect differences in the orientation of the developmental transcription of the yellow locus in Drosophila mela- SUPor-P transposon relative to the end of the chromo- nogaster. Genes Dev. 1: 996-1004. GEYER, P.K., and V. G. CORCES,1992 DNA position-specific repres- some. For example, if the SUPor-P transposon was ori- sion of transcription by a Drosophila zinc finger protein. Genes ented such that yellow was more proximal than white, Dev. 6: 1865-1873. the su(Hw) protein could block the assembly of both GEYER,P. K., C. SPANAand V. G. CORCES,1986 On the molecular mechanism of gypsyinduced mutations at the yellow locus of Dro- genes into arepressive chromatin structure.However, if sophila melanogaster. EMBO J. 5: 2657-2662. the yellow gene was more distal, the degreeof repression GEYER,P. R, K. L. RICHARDSON, V. G. CORCESand M. M. GREEN, would be unaffected by su(Hw) mutations. A second 1988 Genetic instability in Drosophila mlanogaster: P-element mutagenesis by . Proc. Natl. Acad. Sci. USA 85: possible explanation for effects of the allelic state of 6455-6459. su(Hw) on yellow expression is that association of the HARRISON, D. A,, M. A. MORTONand V. G. CORCES,1992 The RNA su(Hw) protein with SUPor-Pcouldalter telomeric chro- polymerase I1 15-kilodalton subunit is essential for viability in Drosophila mlanogmter. Mol. Cell. Biol. 12 928-935. matin such that the yellow gene is subjected to fewer HARRISON,D. A., D. A. GDULA,R. S. COYNEand V. G. CORCES,1993 negative influences than in the absence of this protein. A leucine zipper domain of the suppressor of Hairy-wing protein Determination of theorientation of the telomeric mediates its repressive effect on enhancer function. Genes Dev. 7: 1966-1978. SUPor-P elements will provide insights into this ques- HAZELRIGG, T. R., R. LEVIS and G. M. RUBIN,1984 Transformation tion. of white locus DNA in Drosophila: dosage compensation, zeste In summary, mobilization of a P element containing interaction, and positioneffects. Cell 36: 469-481. HENIKOFF,S., 1990 Position-effect variegation after 60 years. Trends su(Hw) binding regions provides the opportunityto pro- Genet. 6: 422-426. duce anew spectrum of P-induced alleles. Our findings HOLDRIDGE,C., and D.DORSETT, 1991 Repression of' hsp70 heat demonstratethat mutations can be generatedat a shock gene transcription by the suppressm ofHaily-wing protein of Drosophila mlanogmter. Mol. Cell. Biol. 11: 1894-1900. higher frequencyusing this composite transposable ele- JACK,J., D.DORSETT, Y. DELOTTOand S. LIU, 1991 Expression of ment and that a subset of these mutations are depen- the cut locus in the Drosophila wing margin is required for cell dent solely on the presence of the su(Hw) protein. Fur- type specification and is regulated by a distal enhancer. Develop- thermore, SUPor-P insertions can be recovered in ment 113: 735-748. KARPEN, G. H., 1994 Position-effect variegation and the new biology regions of thechromosome normally repressive to of heterochromatin. Curr. Opin. Genet. Dev. 4 281-291. marker gene expression due to insulation of position KARPEN,G. H., and A. C. SPRADLINC,1992 Analysis of subtelomeric heterochromatin in the Drosophila minichromosome Dp1187 effects by binding of su(Hw) protein. by single P elementinsertional mutagenesis.Genetics 132: 737-753. We thank JOHNG LIM for his assistance in the in situ localization KASSIS, J. A.,E. NOLL, E. P. VANSICKLE,W. F. ODENWALDand N. studies. We thank DANIEL. MALLIN,SCOTT PATTON,JAN PETTIS,KRIS PERRIMON,1992 Altering the insertional specificity of a Dro- TIN SCOTT and JEFF SWAN for their help in the mapping studies and sophilatransposable element. Proc. Natl. Acad. Sci. USA 89: stimulating discussions. We thank JERRY BFACH for dependable CUI- 1919-1923. 1074 R. R. Roseman et al.

KF,I.I.w, M. R., S. loo, R. L. BERGand M. W. YOUNG, 1987 Restric- uously from the telomere and are defined by promoter distance tion of Pvlement insertions at the Notch locus of Drosophila mela- and strength, and by SIR3 dosage. Genes Dev. 7: 1133- 1145. nogaster. Mol. Cell. Biol. 7: 1545-1548. ROBERTSON,H. M., C. R. PRESTON,R. W. PHIILIS,D. JOHNSON- K~WELI,,M. G., 1986 P-M mutagenesis, pp. 59-82 in Drosophila; A S<:HI.IT%,W. K. BENZet ul., 1988 A stable source of P element Practical Apprud~,edited by D. B. R~REK~S.IRL Press, Washing- transposase. Genetics 118: 461-470. ton, DC. ROSEMAN,R. R., V. PIRR~?-I’Aand P. K GLX,R, 1993 The su(H7u) I.E\,ls, R., T. HAZELRIG(;and G. M. RURIN,I985 Effects of genomic proteiu insulates expression of the Drosuphila melanogastcr whitr position on theexpression of transduced copies of the whitegene gene fromchromosomal position-effects. EMRO J. 12: 435-442. of Drosophila. Science 229 558-561. RLWN, G. M., 1983 Dispersed repetitive DNAs in Drosophila, pp. IXvI, J. K., 1993 In situ hybridization with biotinylated DNA. Dros. 329-361 in Mobilv Gmetic Elvmmts, edited by J. A. SHAPIRO.Aca- Inf. Sen,. 72: 73-76. demic Press, New York. LINDSIXY,D. I.., and G. G. ZIMM, 19% Thr Genome of Drosophila RtIBlN, G. M., and A. C. SPRAI)I.ING,1982 Genetictransformation melunogaster. Academic Press, New York. of Drosophila with transposable element vectors. Science 218: Loc:lie, J., M. A. KOTARSKI and K D. TARTOFF,1988 Dosage-depen- 348-353. dent modifiers of position effect variegation in Drosophila and a RUBIN,G. M., Kmwk:~.~.,M. G. and BINGHAM,P. M., 1982 The molec- mass action model that explains their effect. Genetics 120: 181- ular basis of P-M hybrid dysgenesis: the nature of induced muta- 198. tions. Cell 29: 987-994. MAKI.OR,R. L., S. M. PARKHURSTand V. G. CORCES,I986 The Dm SEGAIAT,L., R. PERICIWN,J.-P. Boulu, and J.A.IxwsmI, 1992 The sqbhikz mehogaster gypsy transposable element encodes putative Drosophila pourquoi-pas?/wings-clorun zinc finger protein: oocyte gene products homologous to retroviral proteins. Mol. Cell. Biol. nucleus localization and embryonic requirement. Genes Dev. 6: 6: 1129-1134. 1019-1029. M~I)oI.KI.I.,J., W. BENDERand M. MESEI.SON,1983 D. mrlunogastn. SI>OFFORD,J. B., 1976 Positioneffect variegation in Drosophila, pp. mutations suppressible by the suppressor of Huily-wing are inser- 955-1018 in The Genetzcs and Biology oJDrosophila, edited by M. tions of a 7.3 kb mobile element. Proc. Natl. Acad. Sci. USA 80: ASHBURNER andE. NOWTISKI. Academic Press, London. 1678-1682. TOWER,J., G. H. KARPEN, N. Crw~andA. C. SPRZDIJNG,1993 Prefer- O’HARE,K., and G. M. RUBIN,1983 Structures of P transposable ential transposition of Drosophila Pelements to nearby chromo- elements and theirsites of insertion and excision in the Drosophila somal sites. Genetics 133: 347-359. mrlanogaster genome. Cell 34: 25-35. WAKIMOIO,B. T., and M. G. HWRN, 1990 The effects of chromo- O’KANE, C.J., andW. J. GWIRING,1987 Detection in situ of genomic some rearrangementson theexpression of heterochromatic regulatory elements in Drosophila. Proc. Natl. Acad. Sci. USA genes in chromosome2L of Drusuphila mrlanogastrr. Genetics 125: 84 9123-9127. 141-154. PAKKHURST,S. M., D. A. HARRISON,M. P. REMIN(;ION, C. SPANA, WAI.I.RATH,I,. I.., and S. C. R. ELGIN,I995 Position effect variegation in Drosophila is associated with an altered chromatin structure. R. I.. I~..I.I.~:Yet al., 1988 The Drosophila su(Hw) gene, which controls the phenotypic effect of the ops~transposable element, Genes Dev. 9: 1263-1277. encodes a putative DNA binding protein. Genes Dev. 2: 1205- WII.SON, C., R. K. PEARSON,H. J. BELLEN,C. J. O’KANk., U. GROSS- 1215. NIMAUS rt al., 1989 P-element-mediated enhancer detection: an efficient method for isolating and characterizing developmen- PEIFKK,M., and W. BENDER,1986 The anterobithorax and bithorax tally regulated genes in Drosophila. Genes Dev. 3: 1501-1313. mutations of the bithorax complex. EMBO J. 9: 2293-2303. WOODRCFF,R. C., J. L. BI.OUTand J. N. THOMPSON,.JR.1987 Hybrid PIKR~IIA,V. 1988 Vectors for P-mediated transformation in Dro- dysgenesis in D. mrlanogaster is not a general release mechanism sophila, pp. 437-445 in Vector.y: A >Snruqof Molecular Cluning Vrc- for DNA transpositions. Science 237: 1206- 1208. tors nnd Tkir Uses, edited by R. L. RODRI(;IJELand D. T. DEN- ZHAN;, P., and A. C. SPRADI.IN(;,1993 Efficient and dispersed local I IAKlX. Butterworths, Boston. Pelement transposition from Drosophila females. Genetics 133: QIAN,S., B. VARJAVAN~)and V. PIKROTTA, 1992 Molecular analysis 361-373. of the zeste-whiteinteraction reveals a promoter-proximal element ZHANG, P., and A. C. SPRAI)I.IN(;,1994 Insertional mutagenesis 01 essential for distant enhancer-promoter communication. Genet- Drosophila heterochromatin with single P elements. Proc. Natl. ics 131: 79-90. Acad. Sci. USA 91: 3539-3543. RENAIII.I),H., 0. M. ~ARI(:IO,P. I). ZIEKATH,B. I.. BII.I.INGTON, S. K. C:ti.\BlANI et al., 1993 Silent domains are assembled contin- Communicating editor: M. J. SIMMONS